We found Deveugle et al.'s comparison of several techniques for modeling the facies of fluvial-dominated deltaic reservoirs particularly interesting. The paper's main finding confirms the importance of integrating geological concepts into geostatistical modeling to generate realistic reservoir models. Although we think that the paper is quite thorough in its investigation into several geological themes and multidisciplinary integrations, we want to comment on the following two issues. How the statistics of well data should be used in three-dimensional modeling. How to properly calculate and use facies probability.